Video: Beyond Backup: Identity Resilience for the Modern Enterprise | Duration: 4212s | Summary: Beyond Backup: Identity Resilience for the Modern Enterprise | Chapters: Webcast Introduction (40.005s), Speaker Introductions (62.4s), Identity Attack Evolution (243.52s), Identity Attack Vectors (474.05s), Surgical Recovery Options (1402.4701s), Rubrik's Four Pillars (1599.39s), Identity Monitoring & Analysis (2072.93s), Live Product Demo (2362.29s), Q&A and Closing (3368.7751s)
Transcript for "Beyond Backup: Identity Resilience for the Modern Enterprise": Hello, everyone, and welcome to today's webcast, Beyond Backup, Identity Resilience for the Modern Enterprise, sponsored by Rubrik. If during the webcast you have any questions for our presenters, please enter them into the q and a window at any time. Please note that this webcast is being recorded, and a copy will be available for viewing later later this week. And with that, I'd like to hand the webcast over to Matt Bromeli and Kev Johnson. Good morning, good afternoon, and good evening. Welcome, everyone, depending on wherever in the world you are joining us for. It is great to have everyone here. I'm joined today by Kev. Hey. We're actually gonna have a discussion, if you will, of how Rubik does some awesome, awesome things. And I'm joined today by the one, the only, the man, the myth, the legend, Kev Johnson, who, Kev, if you'd like to take a moment to introduce yourself, please. I think the audience would love to know who you are. Sure. Thanks for the intro, Matt. So, yeah, my name is Kev Johnson. I run technical marketing for the identity business here at Rubrik. I've been with Rubrik about five years now. Worked in technical marketing for about eight. Before that, I spent a long time as a sys admin doing active directory stuff, virtualization, and then kind of switched a little bit, pivoted, did a bunch of work in security space before joining, the vendor space. So, yeah, super excited as always to, get to speak to awesome people like yourselves. When you say awesome people, I'm looking over my my shoulder to see who you're talking to, Kev. I'm joined by I'm joined by you today, not the other way around. It's great to have you here. So, yeah, for everyone who's who's joining, we're joined with Kev today from Rubik. My name is Matt Bromley. I'm over here at SANS. I will be, walking you through, Rubik identity resilience today, and I'm actually gonna go back a slide, just to kind of bring this up and and make sure to talk about, kind of what we're doing here. This is very similar to a SAND product review. We're gonna spend some time today getting familiar with, Rupert identity resilience and talk about some of the features, some of the capabilities, what it looks like, how it works. We were lucky enough, actually, it was a little bit I I think, if if I'm not mistaken here and, Kev, keep me honest on the time line. One of the things that I have the benefit of doing here at SANS is actually getting insight into products as they're being created, as they're being developed. And then those of you who are watching and listening, we turn it around into, you know, webinars and papers like this where we get to share information with you. Kev, I think we first touched base, I wanna say, maybe, like, q three last year, if I'm not That feels about right. Yeah. That does feel about right. And it's yeah. You're absolutely right in terms of in terms of the, like, the amount of stuff that you've kind of see go go from, hey. We're building this thing. It's coming out in three weeks' time to, oh, that's that's that's kind of we've had that for ages now. Yeah. Yeah. Even even since we kind of got to the point where we said, okay. We we need to draw a line. This is what we're gonna publish in the in the in the review paper. There's a whole bunch of extra stuff has come since then. So if we get time today, I'd love to show folks some of that stuff as well. Yeah. Absolutely. And I I wanted to bring that up. We will be looking through a live demo, but, I wanted to just reflect. And, Kev, thank you and the the folks that you work with, your colleagues as well, because it's been, definitely a longer process than than normal, but we wanted to capture all the different changes. And I think, folks, that's kind of like maybe without me going into a single slide, that is our, like, introduction here into Rubrik identity resilience. And one of the most interesting parts about the work we've done with Rubrik and with Kevin and his team in getting this set up is when you're talking about identity as a, as a thing that your solution works around and I'm I'm being vague in general in my terms, and I'm gonna get specific in just a moment here. But when we're talking about identity as kind of this nebulous concept, one of the things that I learned from the Rubik team was that it is a constant constant moving target, a constant moving target. 100%. And it's always changing. And, Kev, I I I have in here, like, one of my very first components about identity being the new attack vector and but before I dig into that a little bit, your career, the five years you spent, and I'm sure the years before doing various security, implementations, how have you seen identity kind of morph and change as, like, this, I I I don't even do do I call it, like, attack surface? Is it a vector? Is it an adversarial objective? Yeah. It's it's weird. So, without without actually knowing what what was gonna be in the in these slides, I I put something out on my LinkedIn, I think, maybe a week ago, just talking about, you know, kind of thinking back to much earlier in my career and what we thought of identity was kind of like plumbing in those days. It's like, you know, people people were not really it was just a thing that you set up. You set up your, back in those days, active directory, and you've mostly left it alone. You would do some password resets, but stuff didn't generally change that much. And it was kind of thought to be it wasn't interesting enough for an adversary to go after. And now we see, you know, we we see the likes of Scattered Spider and all of the other, APTs out there that that have got the cool funky names. You know, identity is now something that's super interesting because it's a really easy way that they can sidestep a lot of the hard work that they would previously have had to do, you know, in order to get malware delivered inside an environment and all of that good stuff. They can they they could they could just, you know, they they could they could steal a session token, and, you know, they're they're they're pretty much in. So it's very it's definitely, it's definitely an area that's keeping me busy just keeping up to date with what the, you know, what the latest threats are. And I I I don't I don't wanna I don't wanna sound kind of, like I'm all out for it myself. But it's, yeah, it's it's it's great for me, for my career, just to kind of be in that position. It's I I love being on the the kind of the tip of the spear. So yeah. Yeah. Yeah. No. Absolutely. I I think I think the way you summarize that was actually a good introduction, for folks to think about it. I I I wanna tap into a couple things that Kev said there. What we're doing is we're kind of outlining the problem here that that Rubrik is is dealing with. I'm not gonna say looking to try to deal with. We got a chance to really dig in and see the platform in in in motion, if you will. And, one of the things that I think the team has recognized really well is that identity is this new attack vector for adversaries. But the reason I introduced the question the way that I did, it's it's this weird nebulous thing where it used to just be, like, you know, protect your usernames and passwords, implement MFA, use a strong password policy, rotate every x number of days. Like, Kev, I feel like it was kind of like just a footnote in most security advice was, yeah. Yeah. Yeah. Password management. Yeah. Yeah. Yeah. Whatever. Alright. Let's move on to the other stuff. And what Don't don't put your password on a post it note on your notes, Sarah. Yeah. Basic basic, like, credential management like that, but a lot of it was wrapped around, as long as this as long as the multifactor doesn't get compromised, then you're fine. Yeah. And now to be fair to everyone out there, multifactor authentication as a concept did cut out a significant percentage of threat actors. Right? There was a lot of attacks that needed to adapt and needed to modify and change, and adversaries were on their toes. Now the most elite adversaries who were already implanted and stuff, this is a different story. But for folks who who who, like, very, very bottom of the barrel, like, you know, scraping password dumps and stuff like that, when you think about identity, that's what most people resorted to is, like, my identity is my username and my password and everything. We've got some really basic statistics in here. But, Kev, one thing I wanted to ask you really quick is identity is not username and password anymore. It is not just an an attack vector. I would say it's also the primary gateway for most organizations to get things done. Is that a fair assessment? Yeah. Absolutely. And not not just the primary gateway, but if, you know, if we think about if we think about how we used to deal with business continuity back when I back when I was a sys admin. You know? The the things that we were trying to solve for were, fire, flood, blood, all all of all of that kind of stuff. So, oh, yeah. Someone's someone's taken down the power in my data center. What am I gonna do to make sure my my business can continue? And all all of those things, you know, you can prepare for all of those things. You take backups, you replicate your backups, you have some stuff in the cloud so that you can spin it up. All of that stuff almost doesn't matter. If if all of your systems are fully available apart from identity, your business can't operate. So it's it's a really it's a really effective way, from a threat threat actor's perspective. You know, if they don't wanna have to go through all of the hassle of, hey. Well, you know, we're gonna encrypt all of your data and then try and get some Bitcoin out of you to, to to give to give you the key for that. Mhmm. All they have to do really is compromise identity. Once you've compromised the identity services, if you're then able to, you know, you can do do things like use conditional access policies to lock the legitimate administrators out of the system. So all of a sudden, they're in control. Your admins can't do anything. It's yeah. The business grinds to a halt effectively. Yeah. Yeah. No. A 100%, and it's we we can't go without mentioning, you know, most organizations okay. Okay. Let me let me frame this differently. If you're a Microsoft shop, you're you're utilizing Entra or a hybrid active directory setup. Right? If you're a Google shop, you're utilizing Google for SSO. But where these things branch out, this is the other kind of thing that increases the complexity of these types of attacks here is just how much things have branched out into a, you know, an identity becoming like like a linchpin to access and stuff like that. And it's no longer like, okay. I get a username and password. I access a thing, and then I need another one, and I move around. I log in once. I get that session, and that session carries across so many different things. And I think that adds to the complexity of the statistics I've got up here. 80% of intrusions leverage compromised credentials, probably more of a reflection of how incredibly essential credentials are to attacks these days, more so than, oh, look at the thing that's always here. And and I think some folks are really easy to be like, well, of course, 80% leverage compromised credentials. Like, that that you need credentials to move around. And I look at it the other way, Kev, of, like, if we recognize that there is a common linchpin here, then that's where we should be putting security in place. Right? If I recognize that credentials are essential, and I'm not trying to rhyme all the time here, but I'm doing it off of it. If I recognize that credentials are essential across the board, then, like, that's where I should be implementing my security. I take it as a least common denominator. Identity is gonna be everywhere. 90% of orgs of attacks on critical infrastructure start with identity. So it is not just a objective. It is not just a TTP along the MITRE ATT and CK framework, or I should say a a tactic along the MITRE ATT and CK framework. It is also the entry vector. You know? And then, Kent, we get to this we get to this weird discussion point where it's easy to think of this in the context of accounts, but when usually people think of accounts, they think of humans. Right? You know? And that they and all that kind of stuff. Right? What happens when I take the human out of the equation? 15% of organizations only 15% of organizations, were confident that they were able to prevent nonhuman identity attacks. So, Kev, here's the other side, and I wanna get your quick opinion on this. Here's the other side of identity. It's not just human accounts. We also need identity for things to move around. Absolutely. And, you know, when when we start thinking about nonhuman identities and again, if I kind of hack back to the the bad old days, when I was a sys admin, what you would do is you you would have an application and you create a service account for that application, and then you would do crazy things like say, well, the password for that account never expires because otherwise the system the the service stops working. Mhmm. And, yeah, that was a that was a kind of ugly way of dealing with things. And now you've got things like manage service accounts, but you also have, you know, in the cloud, you have you you've got your service principles by various different from various different IDPs. All of these things are are growing significantly faster than than the number of human identities as well. And then you mentioned, earlier that you were, you know, dealing with the the the AI security event yesterday. If we think about the approach of, you know, everyone wants to deploy agents, those agents have to have some form of identity. It's a nonhuman identity. So those things are being spun up automatically. And when when you have that kind of thing happening, it's very, very easy to to to kind of lose oversight as to what you've actually got in your estate. So you you kind of don't know what you need to protect. And that's you know, if if you if you can't see what you've got to protect, that's a really bad place to be in from a security perspective. Kev, can we rip for a moment? This is not in the slide, but we're gonna have a discussion just because it came out in the past, like, six hours. I'm not sure if you saw it or not. If not, I can give you the quick TLDR. But while we're talking about identity, nonhuman identities, before we start to get into, some of the things that Rubik does and whatnot, a few weeks ago, a month ago, or whatever, Anthropic announces their new, you know, mythos, mythos, mythos, however you wanna call it. I was wondering how long it'd be before this got mentioned. And, and and and, you know, this gets announced. And and and if everyone if we could bring the, discussion discussion here to a head. Right? They announced this massively, capable model, you know, twenty seven year old vulnerabilities in BST, the ability to kind of find zero days and all this kind of stuff and whatnot. We found out this morning. We're recording we're we're we're live here April 22. We found out this morning that a group of folks sitting in a Discord channel, and I'm being vague based on my information, a group of folks in a Discord channel took a look and lined two and two things up together, and we're like, wait a second. This is how the URIs work, and we have a contractor password for the training provider that Anthropic uses. Can we just, like, kind of merge these two knowledgeable things together? Oh, look at that. Wow. And we come to find out that, the model has actually been accessed by unauthorized parties for a period of time simply because of a contractor password leak coupled with an email addressable URI. And now so, Ted, we talked about employee identities. Right? Like, internal identities. We talked about nonhuman identities. But if I can, very currently, real world, extend the discussion to third party identities, Same vector, same concern, but now look at the risk. So, like, it's not the what are your employees connecting to? It's not what nonhuman identities are doing a thing. It's also who's connecting into your network, what third party credentials have access to things that are sensitive inside of it. And, hopefully, everyone gets where Kevin and I are going with this. It just continues a balloon the wider it gets. Yeah. When you when you start looking at things like external ID and customer IAM and the the scale of the problem, you know, if you if you think about IAM for for your organization, you have a set number of of of human identities. You probably have a significantly higher number of nonhuman identities. But then you take a step back and you think, okay. Well, we've got we've got this this, CIAM, program that so all of our customers are in some kind of external identity provider. Mhmm. They're connecting into a whole bunch of things. The scale of that customer base, if your business is doing well, is significantly larger. So, again, it's if you don't have that visibility, yeah, that's a that's a that's a nasty place to be, from a from a security perspective. So, yeah, I I wasn't fully read in on that specific story. I know. But I I can tell you what I'm doing after this webinar. Yeah. And this is one of those things where, like, the beauty of us doing this live is we get the ability to reflect on, some of these some of these things that have come up. We have a question that's come through, which is a very, very, very targeted question, and it's it's it's a little bit I I think so, Simrat, thank you for asking that question. You're specifically asking about, token stealing and theft and stuff like that. Do me a favor. I'm gonna put a pin in that question because I know Kevin and I are gonna talk about this further. So I see your question, but it's one of those things where the answer might be coming or or maybe we'll build up a little bit of a foundation for a better answer. Alright? So so hang tight. So a little bit of of of kind of a review. You know, between Kevin and myself, we've outlined what is kind of this big problem statement here, and that's what I wanted to do in the first, you know, third, if you will, of of today's webcast was get through the the problem and and how big of an issue identity is. It's really easy to just say it, but when you look at the numbers, you look at the statistics, and you break down just where the spread is, it really comes to light of, like, wow. This is just absolutely massive. There's another component to this, though, which is where I wanted to start to give Rubrik a little bit of props, and and it also leads into our discussion of how the, platform does what it does, which is not only is identity a significant attack vector and a significant perimeter problem for a lot of organizations, it's also not that easy to fix. That's the other problem. When you're dealing with active directory, you're talking about domain, multi forests, you're talking about SSOs, integrations. Folks, you're you're and I'm I'm gonna throw out a bunch of terms. If you if you don't get them, that's okay because it's it's either stuff that you deal with or you just you you've never had to deal with it before. Session revocations, ticket revocations, reissuing certificates, sorry, reissuing VPN credentials if need be, multi forest domain password changes, domain recovery. How do you deal with active directory replication between clean and dirty domain? There's just all these things that happen, and every single, like, little bullet point I just dropped in, whether they were technically correct in their flow or not, is sometimes multiple steps, multiple hours, multiple days in order to get done. So as someone who teaches incident response and, Kev, my career has largely been steeped in incident response. The remediation phase of any incident has always been buckle down, grab some coffee, order a pizza because you're you're gonna need it. Right? I've got a few things called out here, an active directory recovery challenge, for example, 20 steps per domain. So if you're in a multi domain environment, you know, good luck. Manual trust recreation, recreation, I should say, error prone, days or weeks of downtime. And the problem is when active directory goes down, everything goes down, especially, depend I should say, depending on how LinkedIn tied you are to that active directory. I'm also setting the conversation up here. Kevin and I are gonna walk through a few use cases, that actually are very specifically related to this. One of my first takeaways and I'm gonna try and create lists as I go through this for all of our attendees here. Those of you watching live, those of you watching the recording, you know, kind of big takeaway number one was one of the first things that I wanted to see Rubrik do was, like, make the remediation side of it a lot easier. And the thing that we don't need here in security, and I'm teeing this up for Kev's success here, One of the things we don't need in security is another alerting source with very little ways to actually fix things. And that is one of the first delineators I wanna draw here between, you know, other solutions and Rubrik, at least reflective of what we've seen Rubrik do, is I was like, hey. If if all this is is another learning source, then let's treat it like another learning source. Like, what's the fidelity of the alerts? But, Kev, was like, no. No. No. I can actually Yeah. Make a lot of these processes shorter. But but, Kev, what are your what are your, like what have you guys seen from, like, how long it's taking to do these things? Yeah. I mean, we've, so we we have we have a team that's it's like an overlay team as part of our customer support called the ransomware response team. And they basically they're a twenty four seven operation, and they get called in when customers get hit by ransomware. Now I'm not really gonna I'm not really gonna start talking about, you know, all the cool stuff we do backup wise that makes it easy easy for customers to recover their data from a ransomware situation. But what you know, there there's a few other things in there. And one of the things that comes up time and time again is, okay. Yeah. We've been on this journey to the cloud, and we kind of thought we were pretty much all in on cloud, but actually, we still have this small on prem footprint. We're still kinda tied to active directory, so we have these active directory objects that we can't just recover into into Entre. So how do we how do we get our business back up and running in this this kind of situation? And, you know, we we've kind we kind of we've we've used the learnings of, you know, the experiences that those that those teams have had to really help kind of shortcut you know, to to orchestrate as much as possible of that recovery. So whether that's the recovery of of hybrid identities, so, you know, you you have these these active directory constructs that then get written into ENTRE, and then there are ENTRE specific attributes there. And, you know, when when it comes to recovery, we have to be conscious of things like, you know, the the the actual u UID is changing. Because if that if that's the case, then in the cloud, you know, Entre is a is a is a flat directory effectively. There's no there's no hierarchy of of organizational units and things like that. But if we start thinking about how things are tied together in ENTRE, it's like, okay. Well, these are all of the, conditional access policy, the applications that are attached to that identity, how they get the privileges that they've got, whether that's through roles or, you know, direct assignments, things like that. So being able to recover in that kind of scenario is is something that comes up a lot. Mhmm. I would I would say the vast majority of the customers that we that we've dealt with in those those kind of situations, they still have fairly hefty, active directory. You know, in in some cases, they are they're further along their cloud journey. In many cases, they are we've got active directory, and we're pretty much only using Entre because we're in m three sixty five. And then when we start thinking about that from a recoverability perspective, it's like if you if you lose your Entre tenant. Mhmm. What what do you do at that point? Do you Yeah. Do you do you just wanna kind of roll it back to a specific point in time, or have you lost have you lost your trust in that specific tenant? Because if you've lost your trust, then you basically have to toss it away and start again. And, you know, how you go about doing that, some some organizations, we see that, you know, they're like, okay. It's cool. We we're just we're just we just got automation. We'll spin everything up again. And and say, okay. Good good luck with that. I've not seen a good example of that going well. But having that capability to take, to take a backup that Rubrik's taken, spin up a brand new entra tenant and just recover into that, that's that's that's been a major benefit for for customers as well. So Yeah. Yeah. There's there's lots of stuff in there, but also it's not just about it's not just about recovery. Right? Because running through running through a recovery of an active directory forest is a even though even though we've orchestrated a whole bunch of it, it's an intensive you know, it it can still be a time consuming process, especially in very, very large environments. And in in many cases, what you actually wanna do is something a little bit more surgical. So, you know, having that visibility into what activities a specific actor has has done in the IDP. So if you have if you get a signal from, let's just throw out some examples. So so if you've got something like a Defender for identity or a CrowdStrike identity security and their ITDR module picks up that, hey. We've detected some suspicious activity from this actor. So this this is this is the user ID here. This is the point in time that we're looking at for this attack. We we can ingest that data. We can we then automatically correlate. Hey. Here's here's all of the things that that specific actor has done in this IDP and, you know, related IDPs. Now you have that list. You know everything. You know exactly what's what's going on. You don't have to go digging in the SIM. You don't need to figure out how to manually remediate that stuff because you just check a box to select the things, you hit revert, and we we can back all of that stuff out. So Mhmm. It's not just about kind of going, hey, this is a worst case scenario where where where basically throwing all the rest of that stuff in the bin. It's it's about, well, is is a surgical recovery the best approach? Or is something where, actually, we do need to toss all all of that away and, you know, we're rolling back two months or whatever. So having options is is really, really important. And I think one of the things that we do really well there is we plug into the existing ecosystem, and, you know, we have we we you mentioned at the beginning that, you know, you didn't you you weren't kind of interested in something that was just another source of alerts and absolutely 100%. But a lot of the context that we can get by mining, by the analysis that we can perform on the backups that we're taking, but also the the live activity tracking that we're doing, That can be super, super useful context that we can provide back to incident responders like yourself. Yeah. Yeah. No. And I think, Kev, what you've done is set me up perfectly as well too to talk through kind of our takeaway from examining in the rubric. So I'm gonna take a few slides here and actually get into, kind of what we saw from it and and also tap into some of the things that Kev brought up as well. And then we we're actually gonna get to see a demo, of of rubric kind of doing its thing. So before we get into it, you know, I I I think we spent the first half of of our of our webcast here getting into, kind of the overall problem statement, the state of the industry, and then dipping into the different areas where where Rubrik is actually successful. I broke these out into four high level areas, and and this is a lot of us going back and forth. This isn't, Matt's classification of Rubik, more of SANS and Rubik working together to figure out the the best approach here. You know, four core pillars here, protect, detect, respond, and recover. Protect and detect are two of the easiest ones that most organizations are able to do. Your identity provider might actually do a lot of this stuff for you. To Kev's point, it's the response and the recover part where you need a niche approach, kind of a, like, what is the best thing to do? What's the best approach to take? You know? I think from a remediation perspective to my incident responders who are watching this, you're gonna remember a day where it was literally like, okay. That trip wire has been crossed. That's now a full enterprise wide password reset. No ifs, ands, or buts about it. We know which account needs to be reset twice in a twenty four hour period, and it was kind of just, you know, a hard coded, we know how this thing goes. These days, I I think what Kev just outlined a little bit earlier was more of a surgical approach towards it because you might not need a full reset, but then also, what's your level of trust? What's you know, how much risk are you willing to accept if an adversary has compromised your entire, identity? You know, I'm gonna say domain, but I don't mean that in a Microsoft sense, but in kind of, like, an identity sense. One of the things and and, Kev, I don't know if this list has changed, but some of the support that we saw from you guys was active directory, Entra, and then Okta. Is that still correct? That is still correct. Yeah. We should have, we should have another announcement, but I can't talk about it for at least I'm I'm gonna say six weeks. So let's let I'll just I'll just zip it now. Folks, as someone who's been in this industry a while, I'm gonna tell you what Kev just said. One of two things is gonna happen. Rubik's either gonna have an out of band announcement about a really cool feature or you need to go visit their booth at Blackhat. One of the two things is gonna is gonna be true here. Alright? But you should you should do the latter anyways, but nonetheless. So on that on that sense, from a product support perspective, I just wanted to make sure this slide was still relevant before I got into it. Rubrik does provide support across these three identity providers. One of the the key things that I wanted to get into first is I actually work with quite a few organizations who have, two or three of these things in play. It's not irregular to see multiple identity providers in. For anyone who's sitting here like, why would you have multiple? Well, it could just be a combination of technologies that you, you know, procured over the years, could be an organic acquisition, could be you're transitioning from one to the other. And a lot of these things, it's it's very possible. And, really, what I want your takeaway to be here is if you are in a multi IDP environment, you are only increasing the complexity. And when you increase complexity, it is a exponential increase in potential attack vector and potential attack surface. I don't know the numerical exponential value, but I will tell you that the more stuff you have from a technology perspective, it is an exponential greater attack surface you have to deal with. So understanding, hey. I'm already in a tricky multi IDP situation. Having support across some of the big providers, including whatever Kev's going to announce in the future is, gonna be a critical, like, position to understand. Not getting into too many specifics here, but active directory, forest attribute level recovery, Microsoft covers conditional access policies, user groups, app registrations. Quick little industry plug for everyone here. Your users and your groups will likely have different authentication requirements than your applications will. Things like elicit consent, grant attacks, and stuff like that are actually areas where adversaries have been able to implement. It's it's a weird way to describe it, but as a high enough viewpoint, a single factor off attack. And then Okta, obviously, with its various configurations and then its different kind of categorizations, users, groups, and and applications. I wanted to double tap really quick into what, Kev talked about from an identity recovery perspective. Another really cool feature that that we saw and we talked about and, Kev, I think you just kind of outlayed this pretty quick. A traditional approach for domain recovery is lengthy. There's a lot of steps to it. You guys have actually orchestrated quite a lot of that and cut it down by, you know, 75%. I'm being very loose with my numbers here, folks. So what I mean by that is, not every single organization is gonna have only 20 steps and not every workflow will have only five steps. Right? But this idea of, reducing the time it takes to get back to a a recovered state or at least a better operational state. Kevin, I know you drilled into this a few minutes ago. Anything additional about kind of, like, clean room isolation or, like, you know, how fast some of these processes you've seen take with various customer environments? Yeah. So, I mean, as as you said, they when when you when you're in incident response mode, one of the one of the things or or the period that tends to be the most stressful and in in many cases, the most long winded is, like, that period of, you know, how how do I go about doing that recovery? And if I if I think about this this from a places I've been in the past, you know, okay. We need we need to recover active directory. Well, the first thing is in most organizations, that is not something that there's any kind of recovery plan for because people just go, well, AD is multi master. It's, you know, it's everywhere. If something goes wrong with the domain controller, we just shoot it. We we deploy a new one. Away we go. And and that works when your problem the the problem you're solving for is availability, but it doesn't help in the cyber scenario. In fact, that highly distributed nature of of things kind of works against you. Mhmm. But knowing when to recover to is is a challenge. So often what we see happen is people will restore something into an IRE, so clean into a clean room. They will then so you've gotta run through that whole 20 mile steps. You you do all of that stuff. You get active directory up and running. It takes you days. You then start running your your security scams using whichever tooling, you're wanting to do to make sure that the adversary isn't there. And in many cases, you chose the wrong point in time. So you just basically rinse and repeat, and you go through that thing over and over again. So, having that ability to do that clean room clean room recovery, you can do the testing as well because you have an isolated environment there that you're using for a clean room. So you can actually build the muscle memory. You can test and make sure that your recoveries are actually gonna work. So you've got that confidence that when you actually do have to do this in anger, it's gonna work. Mhmm. But also, we have some of the some of the cool capabilities built into the platform. So we're we're also every every backup that we take, we're analyzing that backup against a a threat intelligence fee. Now the threat intelligence fee comes from Google threat intel, so the the folks at Mandiant. Mhmm. There's one point eight one point eight million IOCs in there that we're looking at. And that's that's from a kind of file system backup perspective. So there's a very good chance that if you have been compromised, let's say it's a zero day scenario, that that IOC lands in the in the Mandiant threat feed, then lands in your in your on your Rubrik platform. We can then we will then immediately scan all of those backups. You don't have to do anything manually. We're just we're then gonna go, hey. Okay. This is the first point in time that we saw this in your backups. So now you know when that threat actor was first seen in your directory services, which means that you know the next point the the point further back is likely your safest point to recover to. So having having that kind of visibility is super, super important. But at the same time, you know, again, it's it's gonna depend on risk appetite, but you're probably still gonna wanna recover that into a clean room and then poke at it and just make make do your final checks to make sure you are happy before you before you give that the the rubber stamp and say, yes. That's what we're gonna do. But the the whole point here being, we kind of do a lot of the work on the assumption that at some point you will need that information. So we do that work at the point in time that we take the backup. We do the hashing. We do the analysis of, what I get what, you know, what each identity looks like at each specific point in time. So you can see how they change over time. So when you see suspicious drifting configuration, we can flag these things for you, but also, you know, you can track when a change happened and who made a change. So Mhmm. Let's say you end up in a scenario where, let's say you've got something like a Kerberos thing attack goes on. So you've got an a relatively unprivileged account. Adversary compromises that unprivileged account. They go and request a a Kerberos ticket for a for for an account that's got service principal name attached to it. Typically, those are more highly privileged. It's a route to, it's a route to to privilege escalation. We can track all of that through the, you know, through the process for you. Now, again, we're not looking to replace things like Defender or a CrowdStrike or, you know, whichever whichever ITD tooling you're using there. We work in partnership with those teams. But having a unified view across all of your IDPs, all of your identity systems, the other main benefit that we get there is you can you can build policies once and then apply them across those different IDPs. Mhmm. As opposed to some of the other challenges we see is, hey. Well, we've got a point solution for AD. We've got a point solution for Entra, or maybe maybe we're using native capabilities within Entre, and then we've got something else for things like Okta. You end up with this mishmash of policies that kind of maybe meet the security profile that you want to hit, but it's really, really difficult to monitor and make sure that you're you're hitting all of those all of those points as well. Mhmm. Yeah. While you're on it, or, Kev, do you mind really quick? I I I kinda brought slides to back up. Before we jump into our use cases, do you mind spending a minute or two on some of your visibility into the nonhuman identities as well? And I think you've covered kind of the drift and identity patterns, historical perspectives, and how those work, but how have you guys started to tackle the NHI issue? Sure. Yeah. And and from our perspective, so so the way the way Rubrik works, when we when we take a backup of the IDP, we crack the IDP open, and we can then analyze the data that's inside. We generate a thing called UND or user metadata, which is probably a bad choice of terminology because a lot of the a lot of the metadata in there is about computers. It's about NHIs. So from our perspective, NHIs kind of you know, they have they they just look like an identity. We separate them out so that you if you wanna just look look at your NHIs, and I can show you this in a moment when we get to the demo. Mhmm. But we we can we can show you all of those attributes. We can show you, you know, the privileges that those NHIs have got. We can show you how those NHIs have changed over time. We can show you, you know, when the when they were created. We so you can you can monitor life cycle, things like that. But also when things like, secrets that might not have been rotated in line with, with policies. So, again, all of these things should theoretically happen automatically in the background. It's not always the case, unless you've got visibility into what's going on. A lot of organizations tend to tend to make the assumption that because nothing is setting off red lights, everything is fine. And that's a Yeah. It's a fairly dangerous, approach to things. I But, yeah, we do all of that based on snapshot. So, obviously, a snapshot is a point in time. Every time we we do that, we basically capture the state of every identity. Next time we do it, we capture the state of the identity. So now we can draw a graph that shows you how those things have changed over time. Yeah. Awesome. We've got a a couple of use cases here that I'm gonna I'm not gonna run through too quickly, but I wanna go through these and then hand the, reins over to to Kev, to walk us through, the demo and and what he's got lined up because he's been alluding to some really great features. Ransomware recovery, this is one I think probably rings true with everyone who's been in security. You understand, you know, a ransomware compromise is about as bad as it can get. It's a full active directory compromise, data encrypted, group memberships impacted, all sorts of things capable possible there. You know, we've talked very much about Rubrik's response capabilities, getting into real time detection, time series analysis, surgical rollback, really, really easy orchestrated recovery. The ransomware not only takes kind of every avenue of an organizational response, it also requires every avenue of remediation. And I think that's one where, like, you need to have that surgical approach in in what you're doing. The next use case is our insider threat. And this isn't this isn't to breeze through these in in utility here. Because, you know, it's easy it's easy to paint stuff with the brush of ransomware of, like, alright. Well, if you can't access anything in your domain, how would you go about doing it? And it's like, well, you'd restore access. How about insider threats? This is where I've got an adversary I should say adversary, an internal adversary, an internal user abusing what is their, you know, privileged access into the environment. And I think where where Rubrik does a great job and what we saw in our demo was this understanding of how users can start to drift away from typical normal behavior, not just as a user pool, but also individual users themselves. You can start to see, like, people kind of veering off the tracks a little bit of, like, hey. Not only is this person, you know, anomalistic to the organization, they are also anomalistic to their historical pattern. We've got an audit trail in place to kind of capture that. Use case number three was one that I wanted to make sure we hit on. This is where, you know, we we talk deeply about, like, alright. How does this change from a government agency perspective? Well, here, identity becomes super critical because it's likely tied to, access to citizen portals, very, very, you know, critical services and things like that. And this is where I get to mention, you know, Rubrik, also maintains the right levels of compliance, the right levels of certification, or whatnot from a product perspective. So for those of you who are like, oh, I'm in a regulated thing. I can't do this, can't do that. Not the case. Rubric is applicable in this case, and I'll let Kev drill into this further in his demo. And then last but not least is obviously dealing with the nonhuman accounts. How do you go about it? And this is one of the slides where I'll ask everyone here in the audience, don't raise your hand or or or vote back, but I'll ask you just to think to yourself, like, is this a problem you're even starting to answer questions about, let alone even have your arms around, let alone even further even starting to deal with? One of the use cases that I lean on very heavily whenever I I I teach and talk to folks about identity protection is, an incident that that I was involved with that I that I I should say I I incident that I worked where, an an organization, unfortunately, left service accounts open and, was, you know, super easy for, adversaries to get back in because they just weren't protected. The the organization had their hands around employee credentials. They did not have awareness around nonhuman accounts, and that led to being recompromised. And I feel like this is just gonna be a continuous thing for folks. With that said, Kev, I wanna hand the the the screen over to you. Simran, I I see your question still in place. I wanna answer it after, Kev's demo. And, Scott, I saw your question just come in as well. To both of you, let give me a moment. Let me kind of digest through these questions. We're gonna let Kev do his demo, and then when we come back, you know, we should have time to at least answer these two questions. But, Kev, take as long as you'd like to walk through this, obviously, within the next, next, you know, fifteen, sixteen minutes or so. Awesome. If you just bear with me two seconds. You're fine. Take your time. Logging back into an environment because I've just been timed out. So That's okay. It happens. We've all been there, and that's good account protection policies. It's all good. It's all good. Cool. And alright. That looks like it's loading. Awesome. Okay. So let's do I think I might need to stop sharing in order to might. Yeah. Let me see here if I can stop sharing and then hand it over to you. You should be able to take control now. Perfect. So I'm just gonna run through super, super high level overview of, like, what what we do. I've already kinda mentioned, you know, the first thing we we pretty much everything that Rubrik does starts with taking backups. And, normally, I don't really speak about that too much to to a security audience. We don't just do backup based stuff, but we have those backups because we need them as kind of the last line of defense. But the value that we can extract from having those backups means we can do a whole bunch of stuff. So the first thing is we can build that unified unified, inventory of identity. So we can see from here, we've got all of our identities. We can see immediately which ones are privileged through, you know, the various the the the different icons we've got in there. We can filter these down. So if we wanna look just specifically at on try d, for example, and then maybe we wanna start looking at things like enterprise apps. So we can drill into those. We can see we've got this, this privileged app in here. Drill into that. These these kinds of things can be helpful when you're trying to, trying to identify why something isn't working or maybe, you know, you're looking at, you you you've got an ongoing, attack that you're you're you're taking some, taking inventory off. So we can see that this identity is, privileged. It's it's a it's a a an on try the enterprise app. We can take a look then at the API permissions, and we can see that the reason the reason we've got that privilege in there is that we can read and write all files. We also have full control of all sites in there. So that's pulling the API privileges for that specific enterprise app, and and giving us giving us visibility in that. We don't appear to have any activity, so we don't have any changes there for that specific identity, over the last seven days or so. But if I just pick a random, who's a good one? Peter Rabbit's usually a good usually good phone. So Peter Abbott is a human identity. We can see we've got a bunch of violations. I'll cover violations in a moment. But, again, we can see he is privileged. He has a noncompliant password, so he's not meeting the corporate password policy in terms of, you know, maybe that's how long how complex the password is or how regularly that needs to be, updated. But if we look in here oh, come on, Peter. You must have done something. So in the past twelve months, we can see here, Peter's identity has been amended, And it looks like we've just got a bunch of actual log on change log on timestamp changes. So these are legit logins, but this could equally be, you know, it could be some some bad some kind of bad thing. We also have this other this thing up here called Ruby. I'm not gonna do this real time because there's agents behind it, and it can take a couple of moments. So, but, thankfully, if there's any folks out any folks of kind of my era, in The UK, we have the we established this children's TV program called Blue Peter, and they always have one that they prepared earlier. So what I did yesterday, I came in here and I said, okay. Right. Just tell me which identities made the most changes to active directory in the past fourteen days. It goes away, thinks about it. It takes a look at the details that we've got, and it says, actually yeah. So we have we have an account called Peter Abbott. It's made 228 total changes, 737 total activities as well. And we can see the the identities that have made those changes, and we can then kind of drill in and say, okay. Show me what types of changes. So you don't need to be able to drive Rubik in order to be able to get the value out of these things. You can just come in and chat with that. Now I mentioned, I mentioned violations earlier. So, again, this is all based on the analysis of backups. So, again, we take those backups, we crack them up, and we analyze configurations. We then have this this policy engine, that has a bunch of policies that are mapped against various different frameworks. So, you know, your standard kind of MITRE attacks, but also, you know, we've built things like CIS, NIST CSF, ISO 27,001, and then various different kind of actual regulatory frameworks. So we map all of these things. If you if if compliance with these things is something that's important to you, then, you know, we're gonna show you the changes that you need to make. Where we find things that are outside of, you know, that that are not in compliance with these, we're gonna show these as violations. If I just go back to the main dashboard, we have these we got thirty eight thirty seven nearly 38,000 violations in this environment. I'm just gonna pick one that where I know we've got some high severity, so we'll go with identity privileges. If I just grab that first violation there, we can see all of the all of the accounts that we have a delegation enabled for AD privileged identities. Now, obviously, a privileged identity in active directory is, you know, it's a thing. You're gonna see them all the time. But if those identities have got delegation enabled, any any other identity who's got delegated access to that identity can effectively operate. They can do everything that the, that that that identity can do. So let me just find one that hasn't already been, worked on. So we've got LCKD here. I drill into this. So, again, we're giving you a list of stuff that you need to be aware of. So, yeah, we we have delegation enabled. Here's all of the stuff that's that's in the in the policy engine. Here's how you go about fixing it. And, you know, when if if I'd have had this kind of thing back when I was a a a Windows sysadmin, then this would have been great. But the last thing that I want is, hey. Here's a laundry list of work for you to go and do. So we can actually just in here, we could create a ticket. So if you're using something like a ServiceNow, we have integrations there, so we can just, go let's just put a comment in here, sounds webinar. This will then go create a ticket in ServiceNow, and apparently it won't. So I guess someone's taken down that ServiceNow instance. That's, the joy of live demos, folks. There we go. So, yeah, we we can we can create those tickets in our ServiceNow instance. I'll go look at that after the event. But, the other thing that we can do is we can just come in here and fix it. So if you have, you know, if if you have a policy that says, actually, you want your IAM team to be able to see a thing like this and directly fix it, all they have to do is select the domain controller they wanna operate on. And, again, if this was an Entre, an Entre environment, then that Entre environment, you can you can go in and we'll we'll do all of that stuff. It's all API based integration. And so we could see that the remediation, has been has been begun there. The next thing I'll kind of show you real quick is the activity log, and this is kind of where it all starts to come together for me. So every activity that happens in the IDP, whether that is active directory or whether that's in UNTRA or Okta, we're gonna capture that. We store that in Rubrik. So if an adversary gets into your IDP and is able to clear those logs as part of, you know, as part part of their post, post attack planning, then that's fine. We still have all of these details here. Now if we think, you know, maybe maybe we don't maybe there's some of these things that we don't want to happen or, we weren't happy with it. Let's just find a good example. So if we can find a a group policy object. So group policy objects tend to not be changed that often. So when something like that happens, maybe it's not something that you wanna happen. So we have a policy here. We could see we're actually showing here here's here's what exactly what has happened, when that change took place. Now if we're not happy with that, we can just click rollback down there. And that because we have the backups, we know the state of those GPUs over time. So we can just revert that back to the previous version. Also from an activity perspective, let's see if we've got anything in this environment for our sign in logs. So yeah. So we can see we it looks like we've got Manoj signing in a lot of times in the past twenty four hours. So I may maybe you wanna go have a look at, go have a look at his account, but we could see that's in that's in Entre. We we capture deep details such as, you know, the device that that's been signed in from, the location, all of that good stuff. We don't display it in the product because that's potentially PII if it's if it's like a b BYOD, kind of scenario. But it's in there it's there in the database. And then finally, we start looking at alerts. So alerts are where something bad is happening. So, again, Rubrik has a bunch of logic that we use to determine what an alert looks like. You might have seen, the next thing down, which is alert rules, which gives you the ability to effectively build your own alerts. So let's just grab one of these, GPO modification detected. So, if there's a change to a to a group policy object, you can build build a policy there. And when that gets tripped, that's gonna fire an alert, which means we're gonna see something in this dashboard here. The other thing that we can do from an alerts perspective is we can ship them off to other tools. So we have integrations, as I mentioned a couple of times, CrowdStrike identity security, Microsoft Defender for identity, Google SecOps is in there now. I haven't got that one set up in this lab as yet. But every everything's in there so that we can just make direct API calls. We can feed that information into tools like that, but also into tools like, Sentinel, for example. So we have the other way we can handle this the other way we can handle this is we have notification channels, so we can use webhooks to send pretty much anything you like. So we've got what we got in here? We've got a Zapier endpoint. Let's let's imagine I in fact, no. These so these are actual central instances. I guess someone's killed them in this this environment. But all we have to do is come in here and say, hey. Send the alerts that we're interested in. Send them to Sentinel, and Sentinel will then leverage its own, you know, its its own workbooks that are in there. We have a we have a plug in with Sentinel as well, as as all of the other things. So that kind of gives you a super, super high level overview of, you know, the kind of things that we are doing. One of the other things that came up, in the last couple of days quite a few times with with some of the customers who are kind of, like, looking at this is, you know, they they see the activity log and they're like, okay. Cool. Can I just get a report once a week or once a month that shows everything that's happened? And because we've got all of that information here, we can go, yeah. Sure. Of course we can. If what do you want? Do you want it specifically for, maybe you just want it for UltraID, for example? That's great. You can generate those on demand or you can create a schedule, send that to, send that to your audit team. And now you have this you have this audit log of every single action that's taken place in your IDP as you've kind of gone through, the process. We've got about five, six minutes left, so I'm probably gonna come dial back on that stuff now, but that's that's kind of what we have from an identity resilience perspective. Super, super quickly, just to show you recovery, we have we have a forest here. If I wanna recover that forest, it really is a five step wizard. I'm not gonna run through all of this, but, basically, select the point in time that you wanna recover, select the environment that you want to recover to so we can recover to the same environment or into a clean room, and then basically tell us what you want to recover, where you wanna recover it to. Similar kind of thing for Ontra as well. So stand up a brand new tenant, and we can recover your your environment into that. And that at this point in time is oh, one more thing. The the the standard Colombo. So I mentioned we have the CrowdStrike and the, the the, Defender integration. So here we've got some stuff that's coming from CrowdStrike. So we've got this alert that's come in. This is just the alert from CrowdStrike. This is everything that we've extracted based on the information. So we have, the the unique identifier for this specific account. But we know we know the detection time, we know the start time, the end time of this attack. Based on that information, these are the things that have happened, that this specific identity has done to these four identities. So if we believe that this circuit SQL service account has been compromised, all we have to do is come in here, hit revert. We'll make the API call to either active directory or to Entra. And, basically, the the whole idea here is you can just remove the attack of assistance without having to do a full rollback of your identity services. And on that note, I will shut up for a moment, grab some water, and let's get to some questions. Yeah. No. Awesome. Kev, thank you so much. We've got a few minutes left. A couple of questions lined up here that I want to call out. Number one, what are your recommendations to prevent token stealing and replaying on a machine where the threat actors can simply send an email? Appreciate insight on how organizations can close this gap when the token is stolen, MFA is already satisfied, so on and so forth. First off, Simrad, thanks for this, and apologies. You asked this question, like, forty minutes ago. We're we're I told you we'd get to it. I promised you we would. Really quick for everyone here. This specifically deals around with token theft. This is a type of the credential theft. It is, something that that's worth joining into. But, Kev, how do you guys deal with this type of thing if if, you know, if you if you do have a quick answer for us? Yeah. So I'd on on that point, what I would what I would say is, you know, we we we can obviously help you build policies that will monitor, you know, what what your token life cycle is. You know, people have often what we see where we see that kind of attack happening successfully, it's because the token is valid for potentially longer than it necessarily needs to be. So if you can build policies that can kind of help you lock down that thing, then that's gonna that's gonna help you reduce the risk of that happening. In terms of detecting that kind of activity, so again, we we have that activity log so we can see the actions that are taking place. Really, honestly, you're probably not gonna get too many vendors that are gonna put their hands out like this. I would lean more on something like a Defender for Identity or a CrowdStrike Identity Security to detect that sort of, attack happening at that point in time. That's that to me is something that they're excellent at. And we are more about the kind of the detection of the thing that's happening and then, you know, helping helping you to recover from those kinds of things. We're try we're not trying to kind of go butt heads with, with our partners in Microsoft and and CrowdStrike there. Yeah. Absolutely. Second one that came through with the advancement of low code platforms, Microsoft Power Platform, any thoughts around identity management when it's an agent, bot, or service acting on behalf of a user? Scott mentioned this kind of borders between human IDs and nonhuman IDs. If I can insert something really quick here, Scott, I would I would argue API keys probably fall in the same type of category as well. An API key could be created by a human to represent their user account, but, it's not necessarily, them, if you will. It's more about kind of, like, representing the user and what they do, but that can easily be dropped into an automated script and whatnot. But, Kev, I don't know if you wanted to, maybe in the last minute or so, just double check or double tap on how you guys would handle, like, NHIs and maybe this type of situation. Yeah. And so so, again, it's it's really, really important to kind of understand that gap between between NHIs and the identities that that are creating them. And so with with that activity log, we have the excuse me. Sorry. We have we have that information, like, okay. This identity created this API key, for example. So we can we can we can track the relationships between those kinds of things. It's it's it's it's basically just about, okay, well, how do we track all of those things? And, again, if you kind of look in I'm I'm clicking clicking around on my screen now, but, that's that's obviously not super helpful. But yeah. So we we we track all of that information, so you could you can kinda draw that correlation between between those. It is a tricky one Mhmm. But it's really important to have some kind of tooling that understands the relationships between those those things as well. It's similar between kind of like on prem, I'd I'd need to stop saying on prem, but kind of active directory users, then service principles in things like ENTRE that that are directly correlated with them. They they they are sep they are technically separate identities. They're just identities that are related, and you need to understand the relationships between them. 100%. Awesome, Kev. Thank you so much. We are actually right at time, and that also wraps up our questions as well. So perfect timing. Awesome. Everyone, thank you so much for joining us here today. Kev, it's always great when we get a chance to chat. I'm super glad we got a chance to walk through this. Some awesome discussions, great real world examples, and thanks for taking us through the demo.